author's personal copy€¦ · paleoenvironments at east windham (fig. 8). the manorkill...

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Ecology and evolution of Devonian trees in New York, USA

Gregory J. Retallack a,⁎, Chengmin Huang b

a Department of Geological Sciences, University of Oregon, Eugene, Oregon 97403, USAb Department of Environmental Science and Engineering, University of Sichuan, Chengdu, Sichuan 610065, China

a b s t r a c ta r t i c l e i n f o

Article history:Received 16 January 2010Received in revised form 17 September 2010Accepted 29 October 2010Available online 5 November 2010

Keywords:DevonianGivetianNew YorkCatskill GroupPaleosolPaleobotany

The first trees in New York were Middle Devonian (earliest Givetian) cladoxyls (?Duisbergia and Wattieza),with shallow-rootedmanoxylic trunks. Cladoxyl trees in New York thus postdate their latest Emsian evolutionin Spitzbergen. Progymnosperm trees (?Svalbardia and Callixylon–Archaeopteris) appeared in New York later(mid-Givetian) than progymnosperm trees from Spitzbergen (early Givetian). Associated paleosols areevidence that Wattieza formed intertidal to estuarine mangal and Callixylon formed dry riparian woodland.Also from paleosols comes evidence that Wattieza and Callixylon required about 350 mm more mean annualprecipitation than plants of equivalent stature today, that Wattieza tolerated mean annual temperature 7 °Cless than current limits of mangal (20 °C), and Callixylon could tolerate temperatures 14 °C less than modernmangal. Devonian mangal and riparian woodland spread into New York from wetter regions elsewhereduring transient paleoclimatic spikes of very high CO2 (3923±238 ppmv), and subhumid (mean annualprecipitation 730±147 mm) conditions, which were more likely extrinsic atmospheric perturbations ratherthan consequences of tree evolution. For most of theMiddle Devonian CO2was lower (2263±238 ppmv), andpaleoclimate in New York was semiarid (mean annual precipitation 484±147 mm). Such transientperturbations and immigration events may explain the 40 million year gap between the late Emsian(400 Ma) evolution of trees and Famennian (360 Ma) CO2 drawdown and expansion of ice caps.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

The Devonian evolution of trees has been considered responsiblefor a cascade of global carbon sequestration culminating in the LatePaleozoic Ice Age (Berner, 1997; Ward et al., 2006), and perhaps alsoLate Devonian (Frasnian–Famennian) mass extinction (Algeo andScheckler, 1998; Algeo et al., 1995; Ward, 2009). This study of MiddleDevonian paleosols aims at understanding the paleoclimate, includingatmospheric CO2, and paleoenvironment of fossil trees in the CatskillMountains of upper New York State (Banks et al., 1985). Three specificproblems have arisen from past study of these paleosols. First, manyDevonian paleosols now known from New York and Pennsylvania arelike pedocals of semiarid Utah (Driese andMora, 1993; Retallack et al.,2009), too dry to have supported trees. Second, some Devonian treesin New York were intertidal, comparable with modern mangroves(Driese et al., 1997, 2000;Mintz et al., 2010), yet new paleogeographicreconstructions place New York at a paleolatitude of 35°S (J. Golonkafor Joachimski et al., 2002), like Auckland today (37°S) wheremangroves at the extreme southern margin of their mainly tropicaldistribution are only 1 m tall (Taylor, 1983). Third, large paleoclimaticvariations from pedogenic carbonate evidence of varied atmospheric

CO2 levels (Cox et al., 2001) may be coordinated with events ofmarked biotic overturn (Brett et al., 2009). Can such paleoenviron-mental shifts or differing plant climatic ranges explain suchdiscrepancies in the size, distribution and ecology of the earliesttrees to reach North America?

These questions are addressed here with a compilation of data onfossil plants of New York and nearby Pennsylvania (Figs. 1 and 2), andwith detailed studies of a sequence of paleosols in the ManorkillFormation near East Windham (Figs. 3 and 4). Climate, organisms,parent material, paleotopography and time for formation are not onlysoil forming factors, but also ecosystem defining factors (Jenny, 1980).In deep time these various factors can be inferred from paleosols,which represent long-term effects of ecosystems on their substrates(Retallack, 2001). Paleoprecipitation and paleotemperature, forexample, can be interpreted from chemical weathering indices aswell as depth to carbonate in soils (Retallack, 2005a,b; Sheldon, 2006;Sheldon and Tabor, 2009; Sheldon et al., 2002). Isotopic compositionof pedogenic carbonate and depth to carbonate can be used toestimate atmospheric CO2 levels (Cerling, 1991; Ekart et al., 1999;Retallack, 2009a). Degree of chemical weathering can be assessedfrom molar depletion of major elements and strain due to volumechanges (Brimhall et al., 1992; Retallack and Mindszenty, 1994). Thisstudy of Middle Devonian (Givetian) soils from East Windham usessuch proxies to evaluate paleoenvironmental controls on earlycommunities with trees.

Palaeogeography, Palaeoclimatology, Palaeoecology 299 (2011) 110–128

⁎ Corresponding author. Fax: +1 541 346 4692.E-mail address: [email protected] (G.J. Retallack).

0031-0182/$ – see front matter © 2010 Elsevier B.V. All rights reserved.doi:10.1016/j.palaeo.2010.10.040

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

j ourna l homepage: www.e lsev ie r.com/ locate /pa laeo


2. Geological setting

The Manorkill Formation in the Catskill Mountain Front recordsprogradation of a coastal plain westward from uplift of the risingAcadian Mountains to the east. Marine gray siltstones of the PlattekillFormation are overlain by interbedded gray and red paleosols of theManorkill Formation (Fig. 3), and then red paleosols of the OneontaFormation (Sevon and Woodrow, 1985). Dark shales of ManorkillFormation in road cuts measured for this work (34 m in Fig. 4) havebeen correlated with the base of the upper Tully Limestone, using acombination of lithological correlations and fossil ostracods foundwithin these exposures (Johnson and Friedman, 1969; Knox andGordon, 1999). The Tully Limestone elsewhere has a marine fauna(Brett et al., 2009; Heckel, 1973), including ammonoids of thePharciceras amplexum zone (Kirchgasser, 1985; Work et al., 2007),and conodonts of the Polygnathus varcus zone (Davis, 1975; Sparling,1999), indicating late Givetian and 386 Ma in the time scale ofKaufmann (2006).

Vitrinite reflectance from fossil wood at East Windham (35 m inFig. 4) has been used to estimate burial by 6.5 km of overburden(Friedman and Sanders, 1982). This conclusion has been disputed(Levine et al., 1983), but is supported by additional evidence fromillite crystallinity, fluid inclusion homogenization temperatures,oxygen isotopic composition of fracture-filling carbonate cements,and conodont color alteration index (Mora et al., 1998; Sarwar andFriedman, 1995). This alteration has affected the grain size, mineralcomposition, and color of many of these paleosols: gray shales aresilty with recrystallized illite–chlorite and red siltstones are purplewith hematite (Retallack, 1985). Potash metasomatism is plausible

(Mora et al., 1998), but was limited, because maximum K2O observedwas 4.13 wt.% despite common feldspar (Fig. 5). Also affected was thethickness of beds, as indicated by ptygmatic folding of clastic dikes(sand filled mudcracks) in one paleosol (Fig. 3C: 50.6 m in Fig. 4),which average 72±8.4% of their former assumed straightenedthickness. This is comparable with 71% calculated using a standardcompaction formula for calcareous soils (Eq. (1) for Aridisols ofSheldon and Retallack (2001) to obtain original thickness of soil (Bs incm) from thickness of paleosol (Bp in cm) for a likely burial depth (K inkm) of 6.5 km (Sarwar and Friedman, 1995), as follows.

Bs = Bp = −0:62=0:38e 0:17K −1

� �� ð1Þ

3. Materials and methods

A stratigraphic section near East Windham, New York, wasmeasured in detail from Durso Corner overlook (N42.33667°W74.13333°) continuously through large road cuts to the top of theManorkill Formation in August 1995 (Figs. 3 and 4). Different kinds ofpaleosols (pedotypes) were recognized (Tables 1 and 2), expandingan ongoing pedotype nomenclature for Devonian paleosols of NewYork and Pennsylvania (Retallack et al., 2009). Pedotype names referto a particular kind of paleosol, not its location. Thus the Hynerpedotype was named from a type profile at Hyner (Retallack et al.,2009), but comparable paleosols occur also at East Windham (Fig. 4).In addition, a variety of paleonvironmentally significant measureswere taken from the paleosols (Fig. 5): Munsell hue, nodule size,

Fig. 1. Geological map and Devonian paleosol and plant fossil localities examined in upper New York State.

111G.J. Retallack, C. Huang / Palaeogeography, Palaeoclimatology, Palaeoecology 299 (2011) 110–128


depth to calcareous nodules, thickness of paleosol with nodules anddepth of rooting (Retallack, 2001, 2005a). Samples were collectedfrom a representative profile of each pedotype for laboratory analyses(Fig. 5): major and trace element geochemical analysis by XRF, ferrousiron by potassium dichromate titration (both by ALS Chemex ofVancouver, British Columbia), and bulk density determined from theweight suspended in air and then water of a paraffin-coated clod(following Blake and Hartge, 1986). Petrographic thin sections werecut from the same samples, and 500 points counted using a Swiftautomated point counter to determine grain size distribution andmineral composition (Fig. 5). These and other Devonian paleontolog-ical and stratigraphic data (available at website http://www.uoregon.edu/~dogsci/directory/faculty/greg/about) allow comparison withmodern soil classifications (Food and Agriculture Organization,1974; Isbell, 1996; Soil Survey Staff, 2001).

Additional observations and stratigraphic sections were made ofnumerous paleosols elsewhere in New York and Pennsylvania(Retallack, 1985; Retallack et al., 2009). Measurements were madeof the following nodular paleosols (not carbonate beds) sampled forcarbon isotope study by Cox et al. (2001): Dehli (N42.30105°W74.89786°), Toad Hollow (N42.33100° W74.48225°), Guilford(N42.34039° W75.40309°), East Unadilla (N42.35583° W75.25722°),Mount Hayden (N42.38918° W74.24403°), Potter Hollow(N42.43641° W74.24913°), Gulf Creek (N42.33111° W74.137222°)and Kaaterskill Creek (N42.275556° W74.100253°).

Two datasets relating the height of modern trees to modernenvironmental variables were also compiled for comparison withDevonian tree sizes. Height of 100 trees was measured using anOptilogic laser hypsometer at each of 10 sites with a soil profile alsoexposed along a rainfall gradient in western New South Wales,Australia (Table 3). Climatic data for each site was obtained online(www.bom.gov.au/climate/averages/ accessed 11/15/2008). Height

of intertidal fringe (not riverine or basinal) mangrove trees at sites ofvaried mean annual temperature (Table 4) were compiled frompublished ecological studies (Clarke and Hannon, 1967; Lacerda et al.,2002; Lin and Wei, 1983; Pool et al., 1977; Spenceley, 1983; Taylor,1983), with climatic data supplemented from Müller (1982) andonline (www.climate-charts.com accessed 8/17/2009).

Representative fossils collected during this work (Figs. 6 and 7) arearchived in the Condon Collection, Museum of Natural and CulturalHistory, University of Oregon, Eugene. Work included compilation ofpublished data on fossil plant diversity, stratigraphic levels, and size ofstump casts (Table 5: Beck, 1964; Bridge and Nickelsen, 1985; Bridgeand Willis, 1994; Driese and Mora, 1993, 2001; Driese et al., 1997;Elick, 2002; Gordon, 1988; Mintz et al., 2010; White, 1907).

4. Local soil and sedimentary setting

Road cuts along state highway 23 up the Catskill Front to EastWindham have long been favored by university geology excursionsfor their clear exposures of estuarine and fluvial paleochannels,heterolithic cross stratification, and marsh deposits (Fig. 3D–E;Johnson and Friedman, 1969), as well as fossil plants, ostracods andclams (Friedman and Chamberlain, 1995; Friedman and Lundin, 1998;Knox and Gordon, 1999). Information from paleosols (Fig. 3A–B) cannow be added to these lines of evidence for Middle Devonianpaleoenvironments at East Windham (Fig. 8).

The Manorkill Formation is laterally equivalent to the PantherMountain and other marine formations with brachiopods, trilobitesand crinoids to the west (Brett et al., 2009; Heckel, 1973). It formedwithin a coastal plain west of the Acadian Mountains of metamorphicand igneous rocks of New Hampshire and Vermont (Sevon andWoodrow, 1985). This was the source of metamorphic rockfragments, quartz and feldspar, which dominate parent materials of

Fig. 2. Geological sequence and stratigraphic level of Devonian andMississippian fossil plant localities in New York and Pennsylvania compared with a record of paleosol depth to Bkhorizon (a proxy for mean annual precipitation, from Retallack, 2009a), used to correlate these sequences with marine anoxic events and conodont zonation (Kaufmann, 2006).

112 G.J. Retallack, C. Huang / Palaeogeography, Palaeoclimatology, Palaeoecology 299 (2011) 110–128


the paleosols (Fig. 5). It was a flat coastal plain judging from steep cut-banks as found in meandering streams (Fig. 3E), low-angle hetero-lithic cross stratification as found in alluvial levees (Fig. 3D), andcarbonaceous siltstones comparable with marsh deposits (Johnsonand Friedman, 1969). Fossil ostracods may be evidence of limitedmarine influence (Friedman and Lundin, 1998; Knox and Gordon,1999). Small archanodont bivalves are found elsewhere in tidal flat

assemblages with other marine taxa, but the large (up to 22 cm long)shells of Archanodon catskillensis (Fig. 6A) from East Windham appearto have been the earliest known freshwater bivalves (Bridge et al.,1986; Friedman and Chamberlain, 1995). Among modern unionidbivalves, such large sizes are only attained at temperate latitudes(Bauer, 1992) and in large streams, comparable in drainage area withthe Ohio and Mississippi Rivers (van der Schalie, 1938). Furthermore,

Fig. 3. Sedimentary facies and paleosols from East Windham: A, Farwell over Bucktail pedotype at 49.3 m in measured section; B, Windham pedotype at 33 m; C, prominent mudcracks in Gleasonton pedotype at 50.2 m; D, heterolithic cross strata overlying Sproul pedotype paleosols at 34 m; E, channel incision through Farwell, Bucktail and Sproul pedotypesfrom 49 to 44 m. Hammer handle for scale in each panel is 25 cm long.



preservation as molds and casts with corroded umbos (Fig. 6A) isevidence of acidic stream chemistry, and thus little influence of largelakes or shallow marine water.

Gray siltstones associated with paleochannels (marsh deposits ofJohnson and Friedman, 1969) include fossil twigs, but also containcommon carbonaceous root traces, in paleosols here designated theWindham pedotype (Table 1). Such wetland soils are formed by acombination of continuing increments of sedimentation, organiccarbon accumulation and bioturbation by roots and aquatic inverte-

brates. Munsell value of 4 in Windham pedotype paleosols corre-sponds to an organic carbon content of about 2% by weight forAppalachian Devonian shales (Hosterman and Whitlow, 1981), andplant compressions are too sparsely distributed to qualify as peat.Bioturbation is the most prominent of these processes in formation ofthe Windham pedotype, which has little trace of sedimentarybedding. For these same reasons, these paleosols are classified asgleyed Inceptisol (Aquept of Soil Survey Staff, 2001), rather than apeaty soil (Histosol) or alluvial soil (Entisol). Histosols were not found

Fig. 4. A geological section of Devonian paleosols on the Catskill Front east of EastWindham. Scales for degree of development of the individual paleosols and degree of reaction withHCl are after Retallack (2001).



at East Windham, but green-gray Entisols with well preserved fossilplants (Aquents) are here designated the Lookout pedotype (Table 1),after nearby Point Lookout. The unoxidized Windham and Lookoutpedotypes reflect low-lying parts of the coastal plain in which watertable was close to the surface throughout the year. Root penetration ofup to 81 cm (calculated from observed 58 cm using Eq. (1)) in oneWindham pedotype paleosol is evidence that water table fell at leastthat deep during a dry season. Such water table fluctuations alsoexplain the blocky ped structure and large amount of very decayedorganic matter within the Windham pedotype, as also interpreted forcomparable Middle Devonian (Givetian) paleosols from Alberta,Canada (Williams et al., 1996).

Other paleosols are red with hematite and reflect well drainedparts of the landscape, which had a measurable relief of at least 9.2 m(calculated from observed 6.6 m using Eq. (1)) at the paleochannelcut bank (Fig. 3E, 50.6 m in Fig. 4). Formerly good drainage of thesepaleosols also is indicated by deeply penetrating root traces andprobable millipede burrows (Beaconites antarcticus of Gordon, 1988),clay skins in petrographic thin sections, and hydrolytic weatheringrevealed by chemical base loss (Fig. 5). Many of these paleosols(Hyner, Bucktail, Gleasonton and Farwell pedotypes, but not Dursoand Renovo pedotypes) have drab mottles in surficial layers andhaloes around root traces down into the reddish matrix lower in theprofile. Such features are widespread in red paleosols and interpreted

Fig. 5. Petrographic and geochemical data on pedotypes of Devonian paleosols from the East Windham section. Grain size and mineral data is from point counting petrographic thinsections. Molecular weathering ratios are designed as proxy for a variety of weathering reactions.



as burial gleization, a process in which microbial chemical reductionof iron oxides is fueled by consumption of organic matter at thesurface of a soil once it has been freshly buried and subsided to nearthe water table (Retallack, 2001). The alternative explanation ofsurface-water gleization (chemical reduction by surface ponding) isnot favored in this case, because petrographic study (Fig. 5) reveals nodiscontinuities in clay content or porosity, which would create aperched water table. Hyner, Bucktail, Gleasonton and Farwellpedotypes were thus well drained, lowland soils of alluvial leveesand terraces, and stream margins tributary to estuarine and wetlandbasins with Windham and Lookout pedotypes.

With the exception of the Durso pedotype, other red and mottledpedotypes are based on type profiles in the Late Devonian (Famennian)Duncannon Member of the Catskill Formation near Hyner, Pennsylvania(Retallack et al., 2009). These red and mottled pedotypes are differen-tiated by degree of disruption of primary bedding and size and depth ofcarbonatenodules. Sedimentary laminationand ripplemarks are obviousbetween the red-clayey root traces in Renovo and Gleasonton pedotypepaleosols, as in comparable alluvial soils of young sedimentary surfaces(Fluvents of Soil SurveyStaff, 2001). Suchoriginal sedimentary structuresare extensively disrupted by blocky ped structure in Durso and Farwellpedotypes, as in alluvial soils that have been forming for a longer periodof time (Ochrepts of Soil Survey Staff, 2001). Bucktail and Hynerpedotypes also show clear soil structures including a well differentiatedhorizon of calcareous nodules with the calcareous micritic calciasepic-porphyroskelic petrographic texture, replacive caries texture anddisplacive fabrics of pedogenic nodules, as opposed to groundwater or

marinenodules (Retallack, 2001). This horizon is shallower than65 cmintheBucktail pedotype, as indesert soils (Aridisols of Soil SurveyStaff), butdeeper than 65 cm in the Hyner pedotype, which also includeslickensided shearing characteristic of swelling clay soils (Usterts ofSoil Survey Staff, 2001). The calcite of East Windham calcareous nodulesis also isotopically light for carbon (−7.2±1.2‰ vs PDB) and oxygen(−17.7±1.2‰vs PDBanalyses of Coxet al., 2001). A cross-plot of carbonvs oxygen isotopic composition of East Windham pedogenic carbonatesforms a linear trend, as is typical for soils (Knauth et al., 2003), paleosols(Ufnar et al., 2008) and “meteoric diagenesis” (Knauth and Kennedy,2009), but unlike unaltered lacustrine or marine carbonate (Railsbacket al., 2003).

Relative degree of paleosol development can be quantified fromthe diameter (S in mm) of calcareous nodules with the petrographicand physical character of pedogenic nodules, which in modern soils isrelated to soil age (A in ka) by Eq. (2) (with R2=0.57; S.E.=±1.8 kafrom Retallack, 2005a).

A = 3:92dS 0:34 ð2Þ

Application of this equation to East Windham paleosols gives averageduration of soil formation of 11 individual Bucktail paleosols of 4.6±1.8 ka (2.6–7.2 ka), and 6.3±1.8 ka for the single Hyner paleosol atthe top of the measured section (Fig. 4). These were the oldest andbest drained landforms of the Middle Devonian coastal plain. Bycomparison, very weakly developed paleosols (Lookout, Renovo,Gleasonton pedotypes) represented less than 100 yr of soil formation,

Table 1Classification of Devonian pedotypes from East Windham, New York.

Pedotypes Diagnosis U.S soil taxonomy(Soil Survey Staff, 2001)

Food and AgricultureOrganization (1974)

Australian classification(Isbell, 1996)

Bucktail Red siltstone with drab-haloed root traces (A) blocky peds, and shallow(b65 cm) calcareous nodules (Bk) over red siltstone (C)

Calcid Calcic Xerosol Calcic Calcarosol

Durso Red siltstone with red clayey root traces (A) passing down into redsiltstone with blocky peds (Bw) over bedded red siltstone (C)

Ochrept Chromic Cambisol Red-Orthic Tenosol

Farwell Red siltstone with drab haloed root traces (A) over slickensided red siltstone(Bw) and bedded siltstone (C)

Ochrept Chromic Cambisol Red-Orthic Tenosol

Gleasonton Red siltstone with drab-haloed root traces (A) over bedded red siltstone (C) Aquent Eutric Fluvisol Oxyaquic HydrosolHyner Red siltstone with drab-haloed root traces (A) and deep (N65 cm) calcareous

nodules (Bk) over bedded red siltstone (C)Ustert Chromic Vertisol Red Vertosol

Lookout Dark gray shale (O) over green-gray siltstone with carbonaceous root traces(A) over gray siltstone

Fluvent Humic Gleysol Sapric Organosol

Renovo Red siltstone with root traces (A) over bedded red siltstone (C) Fluvent Eutric Fluvisol Stratic RudosolWindham Dark gray shale (O) over green-gray siltstone with carbonaceous root traces

(A) over bedded gray siltstoneAquept Eutric Fluvisol Oxyaquic Hydrosol

Table 2Interpretation of Devonian pedotypes from East Windham, New York.

Pedotype Paleoclimate Paleovegetation Former animals Paleotopography Parent material Time forformation

Bucktail MAT 11.7°±4.4 °C Arid shrubland judging from root traces and profile:no plant fossils

Beaconites antarcticus Dry floodplain Quartzofeld-spathic silts

2–5 kyrMAP 484±147 mmMARP 94±22 mm

Durso No indication Riparian shrubland judging from root traces and profile:Wattieza sp. cf. W. casasii

No trace or bodyfossils

Fluvial levee Quartzofeld-spathic silts

1–2 kyr

Farwell No indication Riparian shrubland judging from root traces and profile:Wattieza sp. cf. W. casasii



0.1–1 kyr

Gleasonton Subhumid warm-temperate

Early successional scrub judging from root traces andprofile: no plant fossils



0.01–0.1 kyr

Hyner MAP 730±147 mm Subhumid woodland judging from root traces and profile:no plant fossils


Dry floodplain Quartzofeld-spathic silts

2–8 kyrMARP 80±22 mm

Lookout No indication Early successional shrubland judging from root traces andprofile: Leclercquia complexa, Tetraxylopteris schmidtii

Archanodoncatskillensis

Oxbow lake margin Quartzofeld-spathic silts

1–2 kyr

Renovo No indication Dry early successional scrub judging from root traces andprofile: no plant fossils


Point bar swalesand chutes

Quartzofeld-spathic silts

0.01–0.1 kyr

Windham No indication Marsh: Wattieza sp. cf. W. casasii No trace or bodyfossils

Oxbow lake margin Quartzofeld-spathic silts

0.1–0.2 kyr



and weakly developed paleosols (Windham, Durso, Farwell pedo-types) represented not much more than 1000 yr. This was a youthfullandscape compared with coasts of plate-tectonic passive margins,such as the modern coastal plain of the southeastern US, where soils500-kyr-old are widespread (Markewich et al., 1990). A bettermodern analog is the Indus river of Pakistan, where soil developmentis limited by high sedimentation rate from Himalayan sources(Ahmad et al., 1977). Summing the durations of paleosols indicatedearlier for the 64 m measured section at East Windham gives a rockaccumulation rate of 0.90 mm yr−1, which can be converted to asediment accumulation rate (using Eq. (1)) of 1.27 mm yr−1. Theseestimates support the idea of synorogenic deposition of the ManorkillFormation (Barrell, 1913; Sevon and Woodrow, 1985), although itsmountainous source was not of Himalayan scale.

5. Paleoprecipitation

Paleosols are archives of former mean annual precipitation andmean annual range of precipitation from depth to, and spread of,calcareous nodules, and from intensity of chemical weathering ofclayeymatrix. Depth to carbonate nodules in soils (Do in cm) increaseswith mean annual precipitation (R in mm following Eq. (3) with

R2=0.52, and standard error±147 mm from Retallack, 2005a).Thickness of soil with carbonate, or distance between lowest andhighest nodule in the profile (Ho in cm), also increases with meanannual range of precipitation, which is the difference in monthlymean precipitation between the wettest and driest month (M in mm:using Eq. (4) with R2=0.58; S.E.=±22 mm recalculated fromRetallack, 2005a).

R = 137:24 + 6:45Do−0:013D2o ð3Þ

M = 0:79Ho + 13:71 ð4Þ

Depths in paleosols could have been compromised by surficial erosionunder paleochannels (Cotter and Driese, 1998) or vertic displacement(Miller et al., 2007; Nordt et al., 2006), but these are not problems forpaleosols of East Windham, where calcic paleosols are remote frompaleochannels and paleosol surfaces are defined by drab-haloed roottraces showing no sign of microrelief formed by vertic displacement(Figs. 3–5). Higher thanmodern atmospheric CO2may increase depth tocarbonate nodules in soils, but an increase from 280 to 3080 ppmvmodeledbyMcFaddenandTinsley (1985) increaseddepth topedogeniccarbonate only 5 cm, so this correction is not made here. However,

Table 3Stations for observations of modern tree height in N.S.W., Australia.

Locality Coordinates Common species Weatherstation

Mean annual precipitation Mean height Tallest plant A horizon thickness Depth to Bk(mm) (m) (m) (cm) (cm)

Gunbar S33.95894 E145.12595 Sclerolaena tricuspis Hay/Hillst. 202.5 0.3±0.1 1.42 16 16Booligal S34.03244 E144.82653 Chenopodium nitrarium Hay/Ivan. 203.9 1.0±0.3 1.54 17 21L. Mungo S33.74500 E143.08167 Atriplex nummularia L. Mungo 225.9 0.3±0.08 0.52 9 23Damara S34.154190 E143.32983 Eucalyptus incrassata Balr./L. M. 273.8 7.4±1.5 10.7 30 35Goolgowi S34.00914 E145.66322 Callitris columellaris Wya./Hills. 299 11.1±2.5 16.1 27 34Balranald S34.55142 E143.58503 Eucalyptus socialis Balranald 321.7 7.5±1.5 11.24 12 37Maude S34.465468 E144.32517 Eucalyptus largiflorens Hay 363.7 10.3±2.1 13.9 30 37Back Cr. S33.86613 E147.35642 Casuarina cristata Wyalong 476 15.7±2.8 19.8 23 65Narrandera S34.75156 E146.50253 Eucalyptus melliodora Narrandera 482.5 14.7±2.6 21.1 17 52Bland Cr. S33.76181 E147.51970 Eucalyptus bridgesiana Marsden 506.5 13.9±3.2 20.9 32 60

Note: Climate data www.bom.gov.au/climate/averages/ accessed 11/15/2008.

Table 4Stations for observations of mangrove height.

Location Coordinates Weatherstation

Mean annual temperature Mangal height Source(°C) (m)

Isla Juan Venado, Nicaragua N12.357853 W87.008156 Managua 27.3 14 Lacerda et al. (2002)Aguirre, Puerto Rico N17.943494 W66.272989 San Juan 26.9 12 Pool et al. (1977)Punta Gorda, Puerto Rico N17.928202 W66.936783 San Juan 26.9 7 Pool et al. (1977)Santa Rosa, Costa Rica N10.851256 W85.794653 Puntarenas 26.9 10 Pool et al. (1977)Laguna de Cosinetas, Venezuela N11.840074 W71.340535 Maracaibo 27.4 7 Lacerda et al. (2002)Morrocoy, Venezuela N10.878871 W68.222421 Maracaibo 27.4 11 Lacerda et al. (2002)Tacarigua, Venezuela N10.536077 W66.107568 Maracaibo 27.4 9.5 Lacerda et al. (2002)Ceiba, Puerto Rico N18.275888 W65.629059 San Juan 26.9 8.5 Pool et al. (1977)Estero Pargo, Mexico N18.743834 W91.550332 Merida 25.8 6 Lacerda et al. (2002)Majana, Cuba N22.686655 W82.793074 Havana 25.1 10 Lacerda et al. (2002)Townsville, Australia S19.272357 E146.835756 Townsville 24.2 8 Spenceley (1983)Marismas, Mexico N22.243437 W97.8000030 Tampico 24.3 5.2 Lacerda et al. (2002)Ten Thousand Islands, Florida N25.788165 W81.424294 Miami 23.9 7.3 Pool et al. (1977)Agua Brava, Mexico N22.175036 W105.616030 Acapulco 27.7 7.5 Lacerda et al. (2002)Lechuguilla, Mexico N25.659523 W109.384900 Mazatlan 24.2 4.5 Lacerda et al. (2002)Sipacate, Guatemala N13.919239 W91.81699 San Salvador 22.8 9 Lacerda et al. (2002)Sepetiba, Brazil S25.055224 W43.935970 Rio de Janeiro 22.7 6.1 Lacerda et al. (2002)Ilha Comprida, Brazil S24.681684 W47.436396 Rio de Janeiro 22.7 8.6 Lacerda et al. (2002)Tiandong, Hui-an, China N25.086640 W118.875029 Jinjiang 20.4 0.65 Lin and Wei (1983)Sucuo, Hui-an, China N24.897865 W118.687102 Jinjiang 20.4 0.8 Lin and Wei (1983)Dongyuan, Hui-an, China N25.016863 W118.942799 Jinjiang 20.4 0.57 Lin and Wei (1983)Xiyuan, Hui-an, China N24.937631 W118.902517 Jinjiang 20.4 0.6 Lin and Wei (1983)Kaitaia, New Zealand S35.000128 E173.260196 Kaitaia 19.6 8 Taylor (1983)Woolooware Bay, Australia S34.004229 E151.156549 Sydney 17.4 6 Clarke and Hannon (1967)Auckland, New Zealand S36.798133 E174.765230 Auckland 15.2 1 Taylor (1983)

Note: Climate data from Müller (1982); Pool et al. (1977), Lin and Wei (1983), and www.climate-charts.com accessed 8/17/2009.



allowance does need to be made for burial compaction of depth to Bkand thickness of soil with nodules, using Eq. (1) described earlier.

Additional indications of paleoprecipitation come from chemicalindex of alteration without potash (A=100mAl2O3/(mAl2O3+mCaO+mMgO+mNa2O), in moles), which increases with meanannual precipitation (R in mm) in modern soils (Eq. (5) withR2=0.72; S.E.=±182 mm from Sheldon et al., 2002).

R = 221e0:0197A ð5Þ

These two distinct methods of obtaining paleoprecipitation areneeded because depth to Bk is highly correlatedwith plant productivityaswell asmean annual precipitation inmodern soils (Retallack, 2009c),and lower productivity vegetation before the evolution of trees mayhave produced shallower Bk horizons than today for the same

precipitation and degree of chemical weathering. Data from paleosolsof theManorkill Formation at EastWindhamdo not support a bias of Bkestimates lower than chemical estimates of paleoprecipitation. The twodifferent estimates of paleoprecipitation are statistically indistinguish-able: 401±147 mm from depth to Bk and 528±182 mm fromchemistry of a Bucktail paleosol at 48.8 m, and 477±147 mm fromdepth to Bk and 642±182 mm from chemistry of another Bucktailpaleosol at 6.3 m (Fig. 4). Middle Devonian paleosols were alsocomparable in weathering with Miocene paleosols of comparablepaleoclimate (Retallack, 1991). Comparability of paleoprecipitationestimates fromchemicalweathering and Bk depth also hold for Silurian,paleosols,with the expected productivity biasfirst noticeable during theOrdovician (Retallack, 1997a, 2008a).

Average estimates of 484±147 mm mean annual precipitationand 94±22 mm mean annual range of precipitation come from 11

Fig. 6. Fossil plants and mollusc from road cuts east of East Windham, New York: A, Archanodon catskillensis from 33.5 m in measured section; B–F, Wattieza sp. cf. W. casasii; B–C,lateral and basal views stem base from 22.3 m; D, deciduous branch from 49.3 m; E–F, stem apex and base in growth position with deciduous branches from 6.6 m; G(to right),Leclercqia complexa partly decorticated leafy stem, from 36.5 m; G–H, Rellimia thomsoni, sporangia (G to left) and branches (H), from 36.5 m. Specimens in Condon Collection,University of Oregon are F112798A (A), F112807 (B–C), F112805 (D), F112795A (E), F112794B (F), F112803A (G), and F112800B (H).



Bucktail pedotype Bk horizons. In contrast, the Hyner pedotypepaleosol at 64 m in the East Windham section has a Bk depthindicative of 730±147 mm mean annual precipitation and Bk spreadindicative of 80±22 mm mean annual range of precipitation. Thiscontrast between common shallow-calcic (Bucktail pedotype) andoccasional deep-calcic (Hyner pedotype) paleosols is characteristic ofDevonian paleosols of Pennsylvania and New York (Fig. 9B). Deep-calcic paleosols such as the Hyner pedotype, occur at irregularintervals that are longer than Milankovitch scale (N100 ka) variationin depth to carbonate (Retallack et al., 2009), and can be used asmarker beds for correlation (Retallack, 2009a,b). These calculationsthus confirm the semiarid nature of most Catskill paleosols, asproposed by Barrell (1913), but not his interpretation that these weremonsoonal soils (mean wet month-dry month precipitation differ-ence of more than 100 mm: Retallack, 2005a). Nevertheless, bothHyner and Bucktail paleosols at Manorkill are evidence of strongerseasonality of precipitation than other Devonian paleosols of NewYork and Pennsylvania (Fig. 9A).

6. Paleotemperature

Deriving paleotemperature from isotopic composition of oxygen(δ18Oc) in soil carbonate (Dworkin et al., 2005) is unreliable for thesepaleosols because of metamorphic alteration of oxygen isotopiccomposition (Mora et al., 1998). Better paleotemperature proxies forpaleosols are alkali index (N=(K2O+Na2O)/Al2O3 as amolar ratio) orchemical index of clayeyness (I=Al2O3/SiO2, as a molar ratio), whichis related to mean annual temperature (T in °C) in modern soils byEqs. (6) and (7) respectively (R2=0.37; S.E.=±4.4 °C for Eq. (6), andR2=0.96; S.E.=±0.6 °C for Eq. (7)). Eq. (6) applies over many kindsofmodern soils (Sheldon et al., 2002), but Eq. (7) is only for Inceptisols(Sheldon, 2006).

T = −18:5 N + 17:3 ð6Þ

T = 46:94I + 3:99 ð7Þ

Both estimates are useful because of potential illitization (Mora et al.,1998), which would effect Eq. (6) but not (7). Low potash and highfeldspar content is evidence against extensive illitization (Fig. 5), and sois the comparable paleotemperature from two Bucktail paleosols

chemically analyzed: 11.7±4.4 °C using Eqs. (6) and 11.8±0.6 °Cusing Eq. (7). These results are evidence that Devonian paleoclimate ofNew York was temperate, not tropical as proposed by Barrell (1913).Thus Bucktail paleosols are evidence of paleotemperature compatiblewith a modern climate at latitude 35°S, which is the paleolatitudeproposed for Late Devonian NewYork by J. Golonka for Joachimski et al.(2002). Hyner paleosols on the other hand, are evidence of unusuallywarm paleoclimates at that paleolatitude (Retallack et al., 2009).

7. Paleo-CO2

Estimates of atmospheric CO2 from Devonian paleosols have beenused to argue for long-term drawdown of atmospheric CO2 after theadvent of woodlands (Cox et al., 2001; Mora et al., 1996; Retallack,1997a), but both climatic and CO2 estimates now reveal considerablevariation, particularly in the early stages of woodland evolution(Fig. 9). No new carbon isotopic analyses are presented here, but theresults of Cox et al. (2001) have been revised for subsequentlydiscovered refinements in estimating atmospheric isotopic composi-tion, temperature-dependent fractionations and soil productivityeffects (Retallack, 2009c).

Paleosol-based estimates of past atmospheric CO2 (Pa in ppmv)come from the balance between isotopically light soil CO2 (δ13Cs)compared with isotopically heavy atmospheric CO2 (δ13Ca) of Eq. (8)(Ekart et al., 1999). A variety of terms in the governing equation arederived from other analyses and transfer functions. The isotopiccomposition of soil CO2 (δ13Cs) comes from that of pedogeniccarbonate (δ13Cc) using temperature (T in °C) dependent fractionationfrom Eq. (9) (R2=0.93, S.E.=±0.28‰: Romanek et al., 1992). Theisotopic composition of associated soil organic matter (δ13Co) is usedas a proxy for respired soil CO2 (δ13Cr) and also to calculate theisotopic composition of atmospheric CO2 (δ13Ca) following Eq. (10)(R2=0.34, S.E.=±0.0039‰: Arens et al., 2000). Finally, partialpressure of respired CO2 in soil (Pr) is derived from its knownrelationship with depth to carbonate nodules (Do, corrected forcompaction in paleosols using Eq. (1)) in modern soils expressed inEq. (11) (R2=0.80, S.E.=±893 ppmv: Retallack, 2009c).

Pa = Pr⋅δ13Cs−1:0044⋅ δ13Cr−4:4

� �δ13Ca−δ13Cs

� ð8Þ

Fig. 7. Paleoclimatic and paleoecological proxies from a measured section of paleosols near East Windham, New York.



δ13Cs =δ13Cc + 100011:98−0:12⋅T

1000 + 1−1000 ð9Þ

δ13Ca =δ13Co + 18:67

� �1:1

ð10Þ

Pr = 66:7Do + 588 ð11Þ

Standard error of estimates of atmospheric CO2 using Eq. (8) wascalculated by Gaussian error propagation of the five componentstandard errors (following Retallack, 2009c). The largest componentof error comes from respired soil CO2 partial pressure (Pr±893ppmv), but propagated error is less than that in proportion toaverages of 10,685 ppmv respired CO2 in 3 individual Hyner paleosols,and 7110 ppmv CO2 in 4 Bucktail paleosols. Of the various carbonisotopic compositions (δ13C) needed for Eq. (8), the isotopiccomposition of carbonate (δ13Cc) was from Cox et al. (2001). Twoother quantities (δ13Ca and δ13Cr) were derived from carbon isotopiccomposition of coexisting organicmatter (δ13Co) taken as−27.2‰ forthe Catskill Formation in Pennsylvania (Mora et al., 1996), which islower than usual for Late Devonian plants (Peters-Kottig et al., 2006).Paleotemperature was calculated using Eq. (6), and paleoproductivityusing Eq. (11), for Bk depths corrected for compaction using Eq. (1).

Atmospheric CO2 levels calculated using these equations from asingle Hyner pedotype paleosol is 3923±415 ppmv, but for 14individual Bucktail paleosols is 2263±238 ppmv. Thus CO2 levelswere generally high but showed excursions of an additional 1700ppmv at times of unusually warm and wet paleoclimate (Fig. 9C). Thisresult is very similar to Late Devonian (366 Ma) paleosols fromHyner,Pennsylvania (Retallack et al., 2009), where atmospheric CO2 levelscalculated using these equations from 3 individual Hyner paleosols

were 3742±320 ppmv, and from 4 Bucktail paleosols were 2009±258 ppmv. These estimates are comparable with those of Cox et al.(2001), and were elevated compared with 700–1275 ppmv calculatedby Mora et al. (1996) for shallow-calcic paleosols of the DuncannonMember of the Catskill Formation. These data support the suppositionof long term drawdown of CO2 in shallow calcic paleosols (Bucktail)fromMiddle Devonian levels of ca. 2200 ppmv to late Devonian levelsof ca. 1000 ppmv based on data of low temporal resolution (Cox et al.,2001; Mora et al., 1996; Retallack, 1997a,b,c). The refined strati-graphic sampling presented here now resolves long-term decline intoa sequence with long episodes of low CO2 punctuated by transientwarm–wet climatic spikes, when CO2 soared to ca. 3800 ppmv. Thesegreenhouse climatic spikes align in time with named marine anoxicevents (Fig. 9), and are also marked in marine rocks by pyritic blackshales and major overturns in invertebrate faunas (Brett et al., 2009).

8. Carbon sequestration

The advent of trees has been considered a boost to carbonsequestration of terrestrial (Berner, 1997; Retallack, 1997a), andmarine ecosystems (Algeo and Scheckler, 1998). Carbon sequestrationin soils was promoted by increased silicate weathering consumingcarbonic acid, and exporting through groundwater to the seabicarbonate and nutrient cations, which promoted carbonate precip-itation (now limestone), algal productivity, anoxia and carbon burial(now black shale) at sea. Carbon burial in peaty soils (now coals) alsoplayed a role, althoughwoody (vitrinite-rich) peats appear only in theLate Devonian (Frasnian: Scheckler, 1986; Retallack et al., 1996), andother kinds of peat such as the Rhynie Chert (Rice et al., 1995) andBarzass Coal (Krassilov, 1981) were formed in marshes of EarlyDevonian herbaceous plants.

Table 5Diameters, heights and paleoclimate of Devonian wood from New York.

Taxon Preservationstyle

No. Locality Reference Stratigraphiclevel

Stem diameterat breast height

PaleosolBk depth

PaleosolAl2O3/SiO2

molarratio

Estimatedtree height

Mean annualprecipitation

Mean annualtemperature

(m) (cm) (cm) (m) (mm) (°C)

Wattieza C 1 Pond Eddy Herein 1066 18 46 4.7±0.9 595±147Wattieza S 1 Wyalusing Elick (2002) 858.6 26.6 75 6.7±0.9 790±147Wattieza S 2 Kaaterskill Bridge and

Nickelsen (1985)661.7 24.5 71 6.2±0.9 757±147

Wattieza S 4 Prattsville Herein 592.3 12 42 3.3±0.9 559±147Wattieza C 5 Manorkill Stein et al. (2007) 336 45 65 10.7±0.9 730±147Wattieza C 4 Windham Herein 320.8 3.3 24 0.2028 1.0±0.9 401±147 13.5±0.6Wattieza C 5 Windham Herein 289.2 1.8 26 0.1386 0.6±0.9 421±147 10.5±0.6Wattieza P 3 Windham Herein 284.9 2 37 0.1386 0.7±0.9 522±147 10.5±0.6Wattieza P 8 Windham Herein 278.7 2.2 32 0.1386 0.7±0.9 477±147 10.5±0.6Wattieza S 61 Gilboa Bridge and Willis

(1994)196.3 36 76 8.8±0.9 793±147

?Duisbergia S 20 W.Saugerties (1) Mintz et al. (2010) −39 6.8 61 2.4±0.9 705±147Callixylon S 1 Newport Driese and Mora

(1993)1232 6 42 0.0840 1.8±0.9 569±147 7.9±0.6

Callixylon S 1 Livingston Gordon (1988) 1162 9 53 2.5±0.9 648±147Callixylon S 1 Trout Run Driese et al. (1997) 1086 30 98 7.4±0.9 884±147Callixylon C 1 Deposit Beck (1964) 1068 15 74 4.0±0.9 775±147Callixylon S 1 Wyalusing Elick (2002) 858.6 14 75 3.8±0.9 790±147Callixylon S 1 Hancock Retallack (1985) 384.1 15 73 4.0±0.9 766±147Callixylon C 1 Hancock Herein 381.4 30 92 0.2072 7.4±0.9 854±147 13.7±0.6Callixylon S 1 Kaaterskill Gordon (1988) 336 15 65 0.2028 4.0±0.9 730±147 13.5±0.6Callixylon S 2 E. Windham Mintz et al. (2010) 336 12.6 65 3.4±0.9 730±147?Svalbardia S 6 W.Saugerties (2) Mintz et al. (2010) 196 10.5 70 2.9±0.9 760±0.9Lepidosigillaria C 1 Naples White (1907) 1068 19 79 4.9±0.9 795±147Lepidosigillaria S 1 Selinsgrove Driese and Mora

(2001)1034 17 74 4.5±0.9 795±147

Note: Preservation styles include stump casts (S), organic log compression (C), inorganic log impression (I), and pyretic permineralized log (P). No. is the number of stumps andstems observed.



The mass transfer of elements in a soil at a given horizon (τw,j inmoles) can be calculated from thebulk density in the soil (ρw in g cm−3)and in the parent material (ρp in g cm−3) and from the chemicalconcentration of the element in soils (Cj,w in weight %) and parentmaterial (Cp,w in weight %), and from the strain indicated by animmobile element in soil (such as Ti used here) compared with parentmaterial (εi,w as a fraction), using Eqs. (12) and (13) (Brimhall et al.,1992; Sheldon and Tabor, 2009).

τj;w =ρw⋅Cj;w

ρp⋅Cj;p

" #εi;w + 1h i

−1 ð12Þ

εi;w =ρp⋅Cj;p

ρw⋅Cj;w

" #−1 ð13Þ

Parent materials of paleosols in the Manorkill Formation aresedimentary rocks derived from erosion of soils, and so are not asideal for this kind of analysis as paleosols developed on crystallinerocks (Retallack and Mindszenty, 1994). Nevertheless, the strain andelemental mass transfer of Devonian paleosols from East Windhamcan be compared with that of dry-woodland paleosols from the

Miocene Dhok Pathan Formation of Pakistan (Retallack, 1991). Thedegree of silicate weathering of Devonian and Miocene alluvialpaleosols is comparable (Fig. 10), and by inference so was produc-tivity and carbon sequestration (Berner, 1997). The analyzedpaleosols from East Windham appear to have supported shrublands(b2 m tall) rather than woodlands (Tables 2 and 5), and chemicalweathering in Devonian and Silurian shrubland precursors of drywoodlands of calcareous soils was comparable with modern aridshrublands (see also Retallack, 1997a). The geographic spread andincreased thickness of woody peats during the Late Devonian(Frasnian to Famennian) is another likely cause of long-term CO2

drawdown and global cooling which culminated in Famennian ice-caps in Brazil (Isaacson et al., 2008), and Tournaisianmontane glaciersin Pennsylvania and Maryland (Brezinski et al., 2008). These eventspostdate the latest Emsian appearance of trees in Spitzbergen(Schweitzer, 1999), and may represent long-term geographic expan-sion of swamps and dry woodlands, despite numerous reversals.

9. Sizes of Devonian trees

Devonian trees are represented in New York by compressed andpermineralized wood and stump casts (Callixylon), sterile and fertile

Fig. 8. A reconstruction of Middle Devonian landscapes, vegetation and soils in New York.



leafy shoots (Archaeopteris: Banks et al., 1985) of progymnosperms,and large compressions of lycopsids (Lepidosigillaria: White, 1907)and cladoxylaleans (Wattieza, including well known “Gilboa trees”:Stein et al., 2007). The very different architecture of these plants mayreflect ecological differences: Callixylon has strongly tapering, deeplypenetrating and spreading basal roots (Meyer-Berthaud et al., 1999),whereas Wattieza has a bulbous basal trunk that is rounded withspreading fine adventitious roots (Stein et al., 2007). Ecologicaldifferences have been investigated using evidence from fossil soils,

which suggest well drained alluvial soils (Psamments, Dystrochrepts,and Usterts) for Callixylon, and poorly drained mangrove soils(Aquents and Sulfaquepts) for Wattieza (Driese et al, 1997; Mintzet al., 2010; Retallack, 1985).

The measured section of paleosols (Fig. 4) at East Windham isstratigraphically above the first known New York occurrence ofcladoxyls such as Wattieza and progymnosperms such as Callixylon–Archaeopteris (Fig. 11). The oldest cladoxyl stumps (West Saugertieslocality 1 of Mintz et al., 2010) are intermediate betweenWattieza and

Fig. 9. Paleoclimatic and pCO2 estimates from Middle (Givetian) to Late (Frasnian) Devonian paleosols of the Catskill Mountains, New York.

Fig. 10.Mass transport and strain (deviation from origin) within paleosols from EastWindham, New York. Strain in soils (y axis) is usually collapse (negative) of volume, but dilation(positive) is possible with eolian or fluvial additions. Mass transfer (x axis) in soils is generally loss (negative), especially of nutrient elements, but can be gain (positive) of insolubleoxides (such as Al and Fe) is also common.



Callixylon in showing a bowl-shaped base and large roots (as in thecladoxyl Duisbergia Schweitzer and Giesen, 2002), and are from nearthe base of theManorkill Formation, 343 m below the detailed section(Fig. 4). Better known is the cladoxyl Wattieza from the Gilboa fossilforest (Mintz et al., 2010; Stein et al., 2007), at a stratigraphic level of122 m (Johnson and Friedman, 1969) below the detailed section(Fig. 4). Other earlyWattieza stump fields near Gilboa are known from45 m and 105 m below our measured section (Fig. 3). The oldestprogymnosperm stumps in New York are from locality 2 at WestSaugerties (of Mintz et al., 2010), which is at the same stratigraphiclevel as the uppermost Gilboa fossil forest. These stumps more likelyto belonged to Svalbardia than Callixylon–Archaeopteris, consideringknown ranges of these plants (Edwards et al., 2000). The oldestsecurely identified leaves and trunks of Callixylon–Archaeopteris inNew York are from quarries near Durham, Manorkill and SouthMountain, which also contain very large Wattieza sp. cf. W. casasii(Berry, 2000; Stein et al., 2007), 3–7 km northeast of the measuredsection, and at a stratigraphic level roughly equivalent to the Hynerpedotype paleosol at the top of the detailed section (Fig. 4).

The basal Oneonta Formation at South Mountain quarry nearManorkill yielded a complete compression of a cladoxyl tree trunk upto 1.2 m basal diameter and 8 m tall, with diameter at breast height(ca 1.2 m) of 45 cm (Stein et al., 2007). The 8 tree stumps at Gilboameasured by Driese et al. (1997), have basal diameters of 15–55 cm,or 41.5±13 cm. Bridge andWillis (1994) illustrate a tall Gilboa stumpwith diameter at breast height of 36 cm. In the Manorkill tree of Steinet al. (2007) diameter at breast height (1.2 m) above the base ofWattieza was a factor of 0.34±0.08 smaller than the swollen base.Assuming similar proportions for Gilboa trees measured by Drieseet al. (1997) gives diameter at breast height of 2.4–22 cm diameter, or14.8±6 cm. Progymnosperm trunks tapered less from their stumps,as shown by a specimen of Callixylon trifilievi (Snigirevskaya, 1984),with roots converging to a stump 60 cm diameter and a diameter at

breast height of 25 cm: a factor of 0.42 smaller. Diameter at breastheight scales with plant height in living trees, and the followingrelationship (Eq. (14) with R2=0.95 and S.E.=±0.9 m from Niklas,1994) between diameter at breast height (B in m) and tree height (Hin m) for 670 living species of all kinds. Of all the equations providedby Niklas (1994), Eq. (14) best predicted the observed height (a littlemore than 8 m) of the near-complete Manorkill tree (of Stein et al.,2007).

H = 21:9B0:896 ð14Þ

Eq. (14) gives an earliest Givetian cladoxyl at West Saugerties(locality 1) some 2.4 m tall, and mid-Givetian Wattieza at Gilboa some8.8 m tall and progymnosperms at West Saugerties (locality 2) some2.9 m tall (Table 5). Forests aremore than 10 m tall, woodlands 2–10 m,and shrublands less than 2 m(Retallack, 1992). Remarkably, noneof thefossil plants found in the detailed section at East Windham (Fig. 4) wasgreater than 4 cm in diameter, and all are predicted using Eq. (14) tohave been less than 1 m tall (Table 5). These include fossils comparablewith Wattieza sp. cf. W. casasii (Fig. 6B–F) in Durso, Farwell andWindham pedotype paleosols, which formed shrublands. Other shrub-lands are represented by twigs and sporangia comparable with theshrubby progymnosperm Rellimia thomsoni (Fig. 6G–H). Leafy stems ofthe prostrate ground-covering lycopsid Leclercqia complexa (Fig. 6G)were found in a Lookout paleosol at 36.5 m in the measured section(Fig. 4). The big trees did not grow in the pedotypes found at EastWindham, but rather in Sulfaquepts (un-named pedotype of Drieseet al., 1997), Dystrochrepts (Peas Eddy pedotype) and Usterts (Hynerpedotype), which had limited distribution in time and space (Retallack,1985, 2009a). These pedotypes moreover correspond with warm–wetCO2-greenhouse spikes evident from paleosols (Fig. 9) and marineanoxic events (Fig. 2). At other drier, cooler and lower CO2 times,vegetation was a mosaic of shrublands with wetland and riparian pole

Fig. 11. Middle Devonian fossil plant ranges and stem diameters in New York.



woodland (Fig. 8). The advent of cladoxyl trees in New York coincideswith the Kačák black shale event (earliest Givetian). Progymnospermtrees and large Wattieza reappeared during the Pumilio Black Shaleevent (mid-Givetian), at a time of a warm wet greenhouse spikeencouraging their spread from elsewhere.

Big plants evolved earlier in more humid climates elsewhere in theworld, such as Germany, where Early Devonian paleosols are commonand non-calcareous (Stets and Schäfer, 2009; Retallack personalobservations). A large cladoxyl (Duisbergia mirabilis) from Wuppertal,in themid-Eifelian (ADMac or velatus-langi zone: Edwards et al., 2000),Brandenberg Schichten, had a diameter at breast height of 23 cm(Schweitzer and Giesen, 2002), corresponding to a height of 5.9±0.9 m(from Eq. (14)). Another large cladoxyl (Protocephalopteris praecox)from the uppermost Emsian to mid-Eifelian, Grey Hoek and WijdeFormations of Spitzbergen, had a diameter at breast height of 30 cm(Schweitzer, 1999), corresponding to a height of 7.5±0.9 m (fromEq. (14)). These are earlier than the first large cladoxyls of earliestGivetian age in New York (Table 5). In Spitzbergen, progymnosperms(Svalbardia polymorpha and its sporeGeminospora lemurata)with stems12.5 cm in diameter, corresponding to plants 3.4±0.9 m tall (fromEq. (14)),first appear in theMimerdalen Formation of earlyGivetian age(predating appearance of Chelinospora concinnus and Cristatisporitestriangulatus: Schweitzer, 1999; Marshall et al., 2007), earlier than mid-Givetian in New York (Table 5). Paleoclimatic limits to the size andimmigration of cladoxyls andprogymnosperms inNewYork cannowbereassessed from their paleosols.

10. Paleoprecipitation constraints on early trees

Modern trees need a minimum amount of rainfall, and theirstature declines steadily with declining mean annual precipitation. Inmany aridlands, the relationship between trees and precipitation is

obscured by intervening sod-grassland ecosystems with their dis-tinctive soils (Mollisols), which evolved since the Miocene (Retallack,2008b), but Australian aridlands retain more ancient kinds of non-mollic soils (Aridisols) in comparable climatic zones (McKenzie et al.,2004). Rather than a patchwork of open and wooded grassland,Australian trees decline in stature to a distinctive kind of polewoodland called “mallee” (Noble and Bradstock, 1989). This decline oftree stature with depth to Bk in paleosols and with mean annualprecipitation has been measured in New South Wales (Fig. 12B,Table 3) for comparison with Devonian trees. Depth to Bk in paleosolswith fossil trees has been measured in New York and Pennsylvania,and these depths can be used to estimate former mean annualprecipitation (using Eq. (3)), once allowances have been made forburial compaction of the paleosols (using Eq. (1)).

The decline in stature of trees with declining mean annualprecipitation is evident from both modern and Devonian trees(Fig. 12B) with the difference that the ecotone between woodlandand shrubland (2 m height) in modern communities is predicted atthe 250±46 mm isohyet, whereas for Devonian Wattieza thisvegetation break was at 571±72 mm and for Callixylon at 611±46 mm mean annual precipitation. These standard errors from aninverse regression to that shown in Fig. 12B are evidence thatCallixylon and Wattieza were not statistically different in theirtolerance of aridity. Devonian trees (N2 m high) required at least ca.350 mm more mean annual precipitation than modern trees. Inaddition, Devonian trees gained stature less rapidly with increasedprecipitation, so that the ecotone betweenwoodland and forest (10 mheight) in modern communities is predicted at the 361±46 mmisohyet, whereas for Devonian Wattieza this vegetation break was at761±72 mm and for Callixylon at 943±46 mm mean annualprecipitation. Only two examples of Lepidosigillaria are known forthis analysis, and both plot with Callixylon (Fig. 12B).

Fig. 12. Correlations between observed stem diameters of Devonian trees and depth to Bk horizon (A) and molar ratio of alumina/silica in paleosols (C), and inferred relationshipbetween tree height and mean annual precipitation (B) and between mangrove height and mean annual temperature (D) during the Devonian compared with today.



These results lend support to the view that Devonian plants werelimited in their colonization of dry soils compared with modern seedplants because they reproduced with spores and so needed waterduring reproduction on exposed ephemeral gametophytes (Niklas,1997). This does not mean to say that Early Devonian and older plantswere restricted to waterlogged soils, where they are best preserved(DiMichele and Hook, 1992): a well known taphonomic artefact forfossil plants of all ages (Retallack, 1998). Aridlands are often awashwith floodwaters (Westbrooke and Florentine, 2005) and support awide variety of pteridophytes, lichens and cyanobacteria (Noble andBradstock, 1989). Devonian trees in well drained, calcareous soils(Aridisols) receiving as little as 570 mm mean annual precipitationare unlikely to have been limited completely by reproduction.Another explanation for smaller than modern Devonian trees couldbe moisture-limited nutrient acquisition by soil microbiota, butcomparison of Devonian and Miocene paleosols does not showsignificant differences in weathering of nutrient cations (Fig. 10). Amore promising explanation is the documented reduction in the ratioof stelar surface area to stelar volume in Silurian and Devonian plants(Niklas, 1997), which is particularly marked in cladoxyls (narrowsteles of Xenocladia compared with wider steles of Pseudosporochnusand Wattieza: Stein and Hueber, 1989) and progymnosperms (lobedstele of Rellimia compared with cylindrical Callixylon: Matten andBanks, 1967). These changes increased transpiration efficiency toincreasingly elaborate plagiotropic photosynthetic shoots (Niklas,1997), but remained less efficient than for plants with large thickleaves appearing during the Mississippian (Beerling et al., 2001).

11. Paleotemperature constraints on early trees

Modern mangroves (Avicennia, Rhizophora, and Kandelia) form tallforests in tropical regions, but bushes only 1 m high at higherlatitudes, such as Auckland harbor, New Zealand (Taylor, 1983) andHui'an, China (Lin and Wei, 1983). At latitudes higher than 37°,intertidal vegetation is entirely herbaceous, salt marsh, including saltgrasses (Spartina and Distichlis) and saltworts (Arthrocnemium andSalicornia: Chapman, 1960). In order to investigate whether thistemperature control acted on the Devonian mangal plant Wattieza(Bridge and Willis, 1994; Driese et al., 1997), data was compiled onmodern intertidal mangroves (Table 3) and the likely mean annualtemperatures (using Eq. (7)) of Devonian trees from associatedpaleosols (Table 5).

Temperature constraints on modern mangroves and Devoniantrees are broadly similar, but the minimal mean annual temperaturerequirement of each is different. Inverse regression (to Fig. 12D)predicts minimum mean annual temperature of 20.2±2.6 °C formodern mangroves, which is similar to their limit in China (20.4 °C),but not New Zealand (15.2 °C). Minimum mean annual temperatureof 13.1±0.5 °C is predicted forWattieza and 6.2±2.6 °C predicted forCallixylon.

The lower temperature limit for Callixylon is unsurprising, becauseCallixylon has never been found in intertidal paleosols and so shouldnot be under comparable constraints. Even within mangroves, treesare larger in inland, low-salinity groundwaters, such as the 39 m tallriparian mangroves (Avicennia germinans) of Belém, Brazil, whereintertidal fringing mangroves are only 9 m tall (Lacerda et al., 2002).

The broadly comparable temperature limits of Wattieza and livingmangroves is surprising because of their very different architecture andreproduction.Wattiezawas similar to fossil lycopsid mangroves such asPleuromeia hataii (Retallack, 1997b). Both genera had a columnar habitlike living mangrove palm (Nypa fruticans) and a pteridophyticreproductive system like the mangrove fern (Acrostichum aureum:Duke, 1992). Most living mangroves have a different branching woodygrowth habit and large seeds, comparable with Carboniferous cordai-tean mangroves, such as Nucellangium glabrum and Cardiocarpusspinatus (Raymond and Phillips, 1983), or Cretaceous angiosperm

mangroves, such as Pabiana variabilis (Retallack and Dilcher, 1986;Upchurch and Dilcher, 1990), although neither of these had viviparouswater-dispersed propagules like many modern mangroves (Duke,1992). These observations suggest that habit and reproductive systemare not temperature-limiting features ofmodern or fossilmangal plants.Amore likely limiting feature is frost sensitivity (Inouye, 2000), becausemany mangrove plants are similar to palms and cacti in havingsucculent stems with large frost-sensitive terminal meristems.

12. Conclusions

The earliest woodland trees in Middle Devonian coastal plains ofNew York were cladoxyl pteridophytes (?Duisbergia). They appearedduring the earliest Givetian at a time of high CO2 and a warmer andwetter climate (Fig. 9), coincident with the Kačák black shale eventand faunal overturn in marine invertebrate communities (Brett et al.,2009). Before and after this transient episode of coastal afforestation,climate was too dry and cool for these trees (Fig. 12) and vegetationconsisted of a mosaic of shrublands and low pole woodlands (Fig. 8).Presumably cladoxyls migrated during warm–wet paleoclimaticspikes fromwarmer and wetter regions such as Germany (Schweitzerand Giesen, 2002) or Spitzbergen (Schweitzer, 1999), wheregeologically older large cladoxyl trees have been found.

Large cladoxyl trees (Wattieza of the “Gilboa forest”) appearedduring themid-Givetian, in sulfidic intertidal soils (Sulfaquept) and soformed mangal vegetation (Driese et al., 1997). Their immigrationcoincided with another CO2-greenhouse spike of warm–wet climate(Fig. 9), at the time of the Pumilio marine anoxic event and faunaloverturn (Brett et al., 2009). At about the same time there alsoappeared in well drained, non-marine soils (Vertisols) the earliestprogymnosperm trees (?Svalbardia) in New York, again later thanearly Givetian progymnosperm trees in Spitzbergen (Schweitzer,1999). Progymnosperms were more limited by precipitation thantemperature (Fig. 12), so may have migrated from higher altitudes orlatitudes. Large Wattieza and Callixylon trees returned duringsubsequent CO2-greenhouse spikes (Geneseo, Genundewa, Middle-sex, Rhinestreet and Kellwasser events of Figs. 2 and 9). Wattiezamangal did not survive the Frasnian–Famennian (Kellwasser) massextinction, but Callixylon–Archaeopteriswoodlands continued into theLate Devonian (Fig. 11), again in unusually warm–wet times andplaces (Retallack et al., 2009).

This complex history may explain the 40 million year lag betweenthe latest Emsian (400 Ma) appearance of trees in Spitzbergen(Schweitzer, 1999) and the late Famennian (360 Ma) Ice Age inAppalachia (Brezinski et al., 2009) and South America (Isaacson et al.,2008), for which trees are considered responsible because ofmitigation of an earlier Devonian CO2 greenhouse by means ofenhanced silicate weathering on land, oceanic fertilization and carbonburial (Berner, 1997). Trees expanded their range during transientCO2-greenhouse crises, but were limited in distribution by temper-ature and precipitation at other times.

The evolution of “killer trees” (Algeo and Scheckler, 1998) or“Medea forests” Ward (2009) has been blamed for Devonian marinemass extinctions, using the argument that enhanced weatheringwould have delivered more nutrients to the ocean and thusencouraged planktic productivity andmarine anoxia. Purely biologicalexpansion of trees would result in a new equilibrium of lower CO2,temperature and precipitation, not multiple transient expansions.Expansions of woodland at times of transient and large increases inatmospheric CO2, paleotemperature and paleoprecipitation suggestanother common cause (Fig. 9). These CO2-greenhouse crises arecomparable with other transient extrinsic perturbations to theatmosphere such as massive volcanic eruptions and meteoriteimpacts (Retallack, 2005b; Retallack and Jahren, 2008; Retallacket al., 2009). These events may have been short lived because newlyevolved woodlands expanded to consume carbon in growth and



weathering, and promote oceanic carbon sequestration, thus restor-ing pre-existing cooler and drier conditions, in contradiction tobiocidal hypotheses (Algeo and Scheckler, 1998; Ward, 2009). By LateDevonian to Mississippian time, trees colonized drier climates andinfertile peaty substrates worldwide (Retallack, 2001), oxygenatingthe atmosphere and initiating the Carboniferous ice age (Berner,1997). This long-term coevolutionary transformation (Retallack,2004) can be contrasted with Milankovitch and other extrinsicvolcanic and bolide perturbations, which were repaired by woodlandexpansion.

Acknowledgements

Tim White, Rudy Slingerland, Douglas Grierson, Dick Beerbower,Conrad Labandeira, Peter Ward and Ted Daeschler offered usefuldiscussion.

Appendix A. Supplementary data

Supplementary data to this article can be found online atdoi:10.1016/j.palaeo.2010.10.040.

References

Ahmad, M., Ryan, J., Paeth, R.C., 1977. Soil development as a function of time in thePunjab river plains of Pakistan. Soil Science Society of America Proceedings 41,1162–1166.

Algeo, T.J., Berner, R.A., Maynard, J.B., Scheckler, S.E., 1995. Late Devonian oceanicanoxic events and biotic crises: “rooted” in the evolution of vascular land plants?GSA Today 5 (3), 63–66.

Algeo, T.J., Scheckler, S.E., 1998. Terrestrial–marine teleconnections in the Devonian;links between the evolution of land plants, weathering processes, and marineanoxic events. Royal Society of London Philosophical Transactions 353, 113–130.

Arens, N.C., Jahren, A.H., Amundson, R., 2000. Can C3 plants faithfully record the carbonisotopic composition of atmospheric carbon dioxide? Paleobiology 26, 137–164.

Banks, H.P., Grierson, J.D., Bonamo, P.M., 1985. The flora of the Catskill clastic wedge. In:Woodrow, D.L., Sevon,W.D. (Eds.), The Catskill Delta: Geological Society of AmericaSpecial Paper, 201, pp. 125–141.

Barrell, J., 1913. The Upper Devonian delta of the Appalachian Geosyncline. AmericanJournal of Science 37, 225–253.

Bauer, G., 1992. Variation in the life span and size of the freshwater pearl mussel.Journal of Animal Ecology 61, 425–436.

Beck, C.B., 1964. Predominance of Archaeopteris in Upper Devonian flora of Catskills andadjacent Pennsylvania. Botanical Gazette 125, 126–128.

Beerling, D.J., Osborne, C.P., Chaloner,W.G., 2001. Evolution of leaf-form in plants linkedto CO2 decline in the Late Paleozoic era. Nature 410, 352–356.

Berner, R.A., 1997. The rise of plants and their effect on weathering and atmosphericCO2. Science 276, 543–546.

Berry, C.M., 2000. A reconsideration of Wattieza Stockmans (here attributed toCladoxylopsida) based on a new species from the Devonian of Venezuela. Reviewof Palaeobotany and Palynology 112, 125–146.

Blake, G.R., Hartge, K.H., 1986. Bulk density. In: Klute, A. (Ed.), Methods of Soil Analysis.Part 1 Physical and Mineralogical Methods: American Society of AgronomyMonograph, 9, pp. 363–375.

Brezinski, D.K., Cecil, C.B., Skema, V.W., Stamm, R., 2008. Late Devonian glacial depositsfrom the eastern United States signal an end of the mid-Paleozoic warm period.Palaeogeography Palaeoclimatology Palaeoecology 268, 143–151.

Bridge, J.S., Nickelsen, B.H., 1985. Reanalysis of the Twilight Park Conglomerate, UpperDevonian Catskill magnafacies, New York State. Northeastern Geology 7, 181–191.

Bridge, J.S., Willis, B.J., 1994. Marine transgressions and regressions recorded in MiddleDevonian shore-zone deposits of the Catskill clastic wedge. Geological Society ofAmerica Bulletin 106, 1440–1458.

Bridge, J.S., Gordon, E.A., Titus, R.C., 1986. Non-marine bivalves and associated burrowsin the Catskill magnafacies (Upper Devonian) of New York State. PalaeogeographyPalaeoclimatology Palaeoecology 55, 65–77.

Brimhall, G.H., Chadwick, O.A., Lewis, C.J., Compston, W., Williams, I.S., Danti, K.J.,Dietrich, W.E., Power, M.E., Hendricks, D., Bratt, J., 1992. Deformational masstransport and invasive processes in soil evolution. Science 255, 695–702.

Brett, C.E., Ivany, L.C., Bartholomew, A.J., Desantis, M.K., Baird, G.C., 2009. Devonianecological-evolutionary subunits in the Appalachian Basin: a revision and a test ofpersistence and discreteness. In: Königshof, P. (Ed.), Devonian Change: CaseStudies in Palaeogeography and Palaeoecology: Geological Society of LondonSpecial Publications, 314, pp. 7–36.

Brezinski, D.K., Cecil, C.B., Skema, V.W., Kertis, C.A., 2009. Evidence for long-termclimate change in Upper Devonian strata of the central Appalachians. Palaeogeo-graphy Palaeoclimatology Palaeoecology 284, 315–325.

Cerling, T.E., 1991. Carbon dioxide in the atmosphere: evidence from Cenozoic andMesozoic paleosols. American Journal of Science 291, 377–400.

Chapman, V.J., 1960. Salt Marshes and Salt Deserts of the World. Interscience, NewYork. 392 pp.

Clarke, L.D., Hannon, N.J., 1967. The mangrove swamp and salt marsh communities ofthe Sydney District: I. Vegetation, soils and climate. Journal of Ecology 55, 753–771.

Cotter, E., Driese, S.G., 1998. Incised valley fills and other evidence of sea-level fluctuationaffecting deposition of the Catskill Formation (Upper Devonian), AppalachianForeland Basin, Pennsylvania. Journal of Sedimentary Research 68, 347–361.

Cox, J.E., Railsback, L.B., Gordon, E.A., 2001. Evidence from Catskill pedogenic carbonatesfor a rapid Late Devonian decrease in atmospheric carbon dioxide concentrations.Northeastern Geology and Environmental Science 23, 91–102.

Davis, W.E., 1975. Significance of conodont distribution in the Tully Limestone(Devonian), New York State. Journal of Paleontology 49, 1097–1104.

DiMichele,W.A., Hook, R.W., 1992. Paleozoic terrestrial ecosystems. In: Behrensmeyer, A.K.,Damuth, J.D., DiMichele, W.A., Potts, R., Sues, H.-D., Wing, S.L. (Eds.), TerrestrialEcosystems Through Time. University of Chicago Press, Chicago, pp. 205–325.

Driese, S.G., Mora, C.L., 1993. Physico-chemical environment of carbonate formation,Devonian vertic palaeosols, central Appalachians, U.S.A. Sedimentology 40, 199–216.

Driese, S.G., Mora, C.I., 2001. Diversification of Siluro-Devonian plant traces in paleosolsand influence on estimates of paleoatmopsheric CO2 levels. Plants Invade the Land:Evolutionary and Environmental Perspectives. Columbia University Press, NewYork, pp. 237–253.

Driese, S.G.,Mora, C.I., Elick, J.M., 1997.Morphologyand taphonomyof root and stumpcastsof the earliest trees (Middle to Late Devonian), Pennsylvania and New York, U.S.A.Palaios 12, 524–537.

Driese, S.G., Mora, C.I., Elick, J.M., 2000. The paleosol record of increased depth ofrooting and changes in pCO2 in the Siluro-Devonian. In: Gastaldo, R.A., DiMichele,W.A. (Eds.), Phanerozoic Terrestrial Ecosystems: Paleontological Society SpecialPapers, 6, pp. 47–61.

Duke, N.C., 1992. Mangrove floristics and biogeography. In: Robertson, A.I., Alongi, D.M.(Eds.), Coastal and Estuarine Studies. American Geophysical Union, WashingtonDC, pp. 63–100.

Dworkin, S.I., Nordt, L., Atchley, S., 2005. Determining terrestrial paleotemperatureusing the oxygen isotopic composition of pedogenic carbonate. Earth and PlanetaryScience Letters 237, 56–68.

Edwards, D., Fairon-Demaret, M., Berry, C.M., 2000. Plant megafossils in Devonianstratigraphy: a progress report. Courier Forschungsinstitut Senckenberg 220, 25–37.

Ekart, D.P., Cerling, T.E., Montañez, I.P., Tabor, N.J., 1999. A 400-million-year carbonisotope record of pedogenic carbonate: implications for paleoatmospheric carbondioxide. American Journal of Science 299, 805–827.

Elick, J.M., 2002. Paleoenvironmental interpretation of a marginal marine environment;the Catskill Formation (Upper Devonian) at Wyalusing Rocks, PA. GuidebookAnnual Field Conference of Pennsylvania Geologists 67, 8–14.

Food and Agriculture Organization, 1974. Soil Map of the World. Volume I, Legend.United Nations Educational Scientific and Cultural Organization, Paris. 59 pp.

Friedman, G.M., Chamberlain, J.A., 1995. Archanodon catskillensis (Vanuxem): fresh-water clams from one of the oldest back-swamp fluvial facies (Upper MiddleDevonian) Catskill Mountains, New York. Northeastern Geology and Environmen-tal Science 17, 431–433.

Friedman, G., Lundin, R.F., 1998. Freshwater ostracodes from Upper Middle Devonianfluvial facies, Catskill Mountains, New York. Journal of Paleontology 72, 485–490.

Friedman, G.M., Sanders, J.E., 1982. Time–temperature-burial significance of Devoniananthracite implies former great (approximately 6.5 km) depth of burial of CatskillMountains, New York. Geology 10, 93–96.

Gordon, E.A., 1988. Body and trace fossils from the Middle-Upper Devonian Catskillmagnafacies, southeastern New York, U.S.A. In: McMillan, N.J., Embry, A.F., Glass, D.F.(Eds.), Devonian of the World, v. 1: Canadian Society of Petroleum Geology Memoir,14, pp. 139–156.

Heckel, P.H., 1973. Nature, origin and significance of the Tully Limestone: an anomalousunit in the Catskill Delta, Devonian of New York. Geological Society of AmericaSpecial Paper 138 244 pp.

Hosterman, J.W., Whitlow, S.I., 1981. Munsell color value as related to organic carbon inDevonian Shale of the Appalachian Basin. American Association of PetroleumGeologists Bulletin 65, 333–335.

Inouye, D.W., 2000. The ecological and evolutionary significance of frost in the contextof climate change. Ecological Letters 3, 457–463.

Isaacson, P.E., Diaz-Martinez, E., Grader, G.W., Kalvoda, J., Babek, O., Devuyst, F.X., 2008.Late Devonian–earliest Mississippian glaciation in Gondwanaland and its biogeo-graphic consequences. Palaeogeography Palaeoclimatology Palaeoecology 268,126–142.

Isbell, R.F., 1996. The Australian Soil Classification. C.S.I.R.O, Collingwood. 143 pp.Jenny, H., 1980. The Soil Resource. Springer, Berlin. 377 pp.Joachimski, M.M., Pancost, R.D., Freeman, K.H., Ostertag-Henning, C., Buggisch, W.,

2002. Carbon isotope geochemistry of the Frasnian–Famennian transition.Palaeogeography Palaeoclimatology Palaeoecology 181, 91–109.

Johnson, K.G., Friedman, G.M., 1969. The Tully clastic correlatives (Upper Devonian ofNew York State: model for recognition of alluvial, dune(?), tidal nearshore (bar andlagoon) and offshore sedimentary environments in a tectonic delta complex.Journal of Sedimentary Petrology 39, 451–485.

Kaufmann, B., 2006. Calibrating the Devonian time scale; a synthesis of U–Pb ID-TIMSages and conodont stratigraphy. Earth-Science Reviews 76, 175–190.

Kirchgasser, W.T., 1985. Ammonoid horizons in the Upper Devonian GeneseeFormation of New York: legacy of the Genesee, Portage, and Chemung. In:Woodrow, D.L., Sevon,W.D. (Eds.), The Catskill Delta: Geological Society of AmericaSpecial Paper, 201, pp. 224–235.

Knauth, L.P., Kennedy, M.J., 2009. The late Precambrian greening of the Earth. Nature460, 728–732.



Knauth, L.P., Brilli, M., Klonowski, S., 2003. Isotope geochemistry of caliche developedon basalt. Geochimica et Cosmochimica Acta 67, 185–195.

Knox, L.W., Gordon, E.A., 1999. Ostracodes as indicators of brackish water environmentsin the Catskill Magnafacies (Devonian) of New York State. PalaeogeographyPalaeoclimatology Palaeoecology 148, 9–22.

Krassilov, V.A., 1981. Orestovia and the origin of vascular plants. Lethaia 14, 235–250.Lacerda, L.D., Conde, J.E., Kjerfve, B., Alvarez-León, R., Alarcón, C., Polania, J., 2002.

American mangroves. In: Lacerda, L.D. (Ed.), Mangrove Ecosystems: Function andManagement. Springer, Berlin, pp. 1–62.

Levine, J.R., Friedman, G.M., Sanders, J.E., 1983. Time–temperature-burial significanceof Devonian anthracite implies former great (approximately 6.5 km) depthof burial of Catskill Mountains, New York: discussion and reply. Geology 11,122–125.

Lin, P., Wei, X.-M., 1983. Ecological notes on themangroves of Fujian China. In: Teas, H.J.(Ed.), Biology and Ecology of Mangroves. Junk, Hague, pp. 33–46.

Markewich, H.W., Pavich, M.J., Buell, G.R., 1990. Contrasting soils and landscapes of thePiedmont and Coastal Plain. Geomorphology 3, 417–447.

Marshall, J.E.A., Astin, T.R., Broan, J.F., Mark-Kurik, E., Lazuaskiene, J., 2007. Recognizingthe Kačák event in the Devonian terrestrial environment and its implications forunderstanding land–sea interactions. In: Becker, R.T., Kirchgasser, W.T. (Eds.),Devonian Events and Correlations: Geological Society of London Publication, 278,pp. 133–155.

Matten, L.C., Banks, H.P., 1967. Relationship between the progymnosperm generaSphenoxylon and Tetraxylopteris. Torrey Botanical Club Bulletin 94, 321–333.

McFadden, L.D., Tinsley, J.C., 1985. Rate and depth of pedogenic-carbonate accumula-tion in soils: formulation and testing of a compartment model. In:Weide, D.L. (Ed.),Soils and Quaternary Geology of the Southwestern United States: GeologicalSociety of America Special Paper, 203, pp. 23–41.

McKenzie, N., Jacquier, D., Isbell, R., Brown, K., 2004. Australian Soils and Landscapes.C.S.I.R.O, Melbourne. 416 pp.

Meyer-Berthaud, B., Scheckler, S.E., Wendt, J., 1999. Archaeopteris is the earliest knownmodern tree. Nature 398, 700–701.

Miller, D.L., Mora, C.I., Driese, S.G., 2007. Isotopically concentric growth in largecarbonate nodules in Vertisols: implications for climate and ecosystem assess-ments. Geoderma 142, 104–111.

Mintz, J.S., Driese, S.G., White, J.D., 2010. Environmental and ecological variability ofMiddle Devonian forests in Appalachian Basin paleosols, New York, United States.Palaios 25, 85–96.

Mora, C.I., Sheldon, B.T., Elliot, W.C., Driese, S.G., 1998. An oxygen isotope study of illiteand calcite in three Appalachian Paleozoic vertic paleosols. Journal of SedimentaryResearch 68, 456–464.

Mora, C.L., Driese, S.G., Colarusso, L.A., 1996. Middle to Late Paleozoic atmospheric CO2

from soil carbonate and organic matter. Science 271, 1105–1107.Müller, M.J., 1982. Selected Climatic Data for a Global Set of Standard Stations for

Vegetation Science. Junk, Hague. 306 pp.Niklas, K.J., 1994. Plant Allometry. University of Chicago Press, Chicago. 395 pp.Niklas, K.J., 1997. The Evolutionary Biology of Plants. University of Chicago Press,

Chicago. 449 pp.Noble, J.C., Bradstock, R.A. (Eds.), 1989. Mediterranean Landscapes in Australia. C.S.I.R.O,

Melbourne. 485 pp.Nordt, L., Orosz, M., Driese, S., Tubbs, J., 2006. Vertisol carbonate properties in relation to

mean annual precipitation: implications for paleoprecipitation estimates. Journalof Geology 114, 501–510.

Peters-Kottig, W., Strauss, H., Kerp, H., 2006. The land plant δ13C record and plantevolution in the Paleozoic. Palaeogeography Palaeoclimatology Palaeoecology 240,237–252.

Pool, D.J., Snedaker, S.C., Lugo, A.E., 1977. Structure of mangrove forests in Florida,Puerto, Mexico and Costa Rica. Biotropica 9, 195–212.

Railsback, I.B., Holland, S.M., Hunter, D.M., Jordan, E.M., Diaz, J.R., Crowe, D.E., 2003.Controls on geochemical expression of subaerial exposure in Ordovician limestonesfrom the Nashville Dome, Tennessee, U.S.A. Journal of Sedimentary Research 73,790–805.

Raymond, A., Phillips, T., 1983. Evidence for anUpper Carboniferousmangrove community.In: Teas, H.J. (Ed.), Biology and Ecology of Mangroves. Junk, Hague, pp. 19–29.

Retallack, G.J., 1985. Fossil soils as grounds for interpreting the advent of large plantsand animals on land. Royal Society of London Philosophical Transactions B309,105–142.

Retallack, G.J., 1991. Miocene Paleosols and Ape Habitats in Pakistan and Kenya. OxfordUniversity Press, New York. 346 pp.

Retallack, G.J., 1992. What to call early plant formations on land. Palaios 7, 508–520.Retallack, G.J., 1997a. Early forest soils and their role in Devonian global change. Science

276, 583–585.Retallack, G.J., 1997b. Earliest Triassic origin of Isoetes and quillwort evolutionary

radiation. Journal of Paleontology 71, 500–521.Retallack, G.J., 1997c. Compaction of Devonian lycopsid stems from the Beacon Heights

Orthoquartzite, southern Victoria Land. Antarctic Journal of the United States 30,42–44.

Retallack, G.J., 1998. Fossil soils and completeness of the rock and fossil record. In:Donovan, S.K., Paul, C.R.C. (Eds.), The Adequacy of the Fossil Record. John Wiley,Chichester, pp. 131–162.

Retallack, G.J., 2001. Soils of the Past. Blackwell, Oxford, Blackwell. 346 pp.Retallack, G.J., 2004. Soils and global change in the carbon cycle over geological time. In:

Holland, H.D., Turekian, K.K. (Eds.), Treatise on Geochemistry. Pergamon, Oxford,pp. 581–605. v. 5.

Retallack, G.J., 2005a. Pedogenic carbonate proxies for amount and seasonality ofprecipitation in paleosols. Geology 33, 333–336.

Retallack, G.J., 2005b. Permian greenhouse crises. In: Lucas, S.G., Ziegler, K.E. (Eds.), TheNonmarine Permian: Bulletin NewMexico Museum of Natural History and Science,30, pp. 256–269.

Retallack, G.J., 2008a. Cambrian paleosols and landscapes of South Australia. AustralianJournal of Earth Sciences 55, 1083–1106.

Retallack, G.J., 2008b. Cenozoic cooling and grassland expansion in Oregon andWashington. Paleobios 28, 89–113.

Retallack, G.J., 2009b. Greenhouse crises of the past 300 million years. GeologicalSociety of America Bulletin 121, 1441–1455.

Retallack, G.J., 2009a. Early Paleozoic pedostratigraphy and global events in Australia.Australian Journal of Earth Sciences 56, 569–584.

Retallack, G.J., 2009c. Refining a pedogenic CO2 paleobarometer for quantifying themiddleMiocene greenhouse spike. Palaeogeography Palaeoclimatology Palaeoecology 281,57–65.

Retallack, G.J., Dilcher, D.L., 1986. Cretaceous angiosperm invasion of North America.Cretaceous Research 7, 227–252.

Retallack, G.J., Jahren, A.H., 2008. Methane release from igneous intrusion of coal duringLate Permian extinction events. Journal of Geology 116, 1–20.

Retallack, G.J., Mindszenty, A., 1994. Well preserved late Precambrian paleosols fromnorthwest Scotland. Journal of Sedimentary Research A64, 264–281.

Retallack, G.J., Veevers, J.J., Morante, R., 1996. Global early Triassic coal gap betweenLate Permian extinction and Middle Triassic recovery of peat-forming plants.Geological Society of America Bulletin 108, 195–207.

Retallack, G.J., Hunt, R.R., White, T.E., 2009. Devonian palaeosols and tetrapod habitatsin Pennsylvania. Geological Society of London Journal 166, 1143–1156.

Rice, C.M., Ashcroft, W.A., Batten, D.J., Boyce, A.J., Caulfield, J.B.D., Fallick, A.E., Hole, M.J.,Jones, E., Pearson, M.J., Rogers, G., Saxton, J.M., Stuart, F.M., Trewin, N.H., Turner, G.,1995. A Devonian auriferous hot spring system, Rhynie, Scotland. GeologicalSociety of London Journal 152, 229–250.

Romanek, C., Grossman, E., Morse, J., 1992. Carbon isotopic fractionation in syntheticaragonite and calcite: effects of temperature and precipitation rate. GeochimicaCosmochimica Acta 56, 416–430.

Sarwar, G., Friedman, G.M., 1995. Post-Devonian Sediment Cover over New York State;Evidence from Fluid-Inclusion, Organic Maturation, Clay Diagenesis and StableIsotope studies. Springer, Berlin. 113 pp.

Scheckler, S.E., 1986. Geology, floristics and paleoecology of Late Devonian coal swampsfrom Appalachian Laurentia (U.S.A.). Annales de la Societe Geologique de Belgique109, 209–222.

Schweitzer, H.-J., 1999. Die Devonfloren Spitzbergens. Palaeontographica B252, 1–122.Schweitzer, H.-J., Giesen, P., 2002. Wichtige neuere Pflanzenfunde im Mitteldevon von

Wuppertal im Bergischen Land (Westdeutschland). Palaeontographica B262,91–117.

Sevon, W.D., Woodrow, D.L., 1985. Middle and Upper Devonian stratigraphy within theAppalachian Basin. In: Woodrow, D.L., Sevon, W.D. (Eds.), The Catskill Delta:Geological Society of America Special Paper, 201, pp. 1–7.

Sheldon, N.D., 2006. Quaternary glacial–interglacial climate cycles in Hawaii. Journal ofGeology 114, 367–376.

Sheldon, N.D., Retallack, G.J., 2001. Equation for compaction of paleosols due to burial.Geology 29, 247–250.

Sheldon, N.D., Retallack, G.J., Tanaka, S., 2002. Geochemical climofunctions from NorthAmerican soils and application to paleosols across the Eocene–Oligocene boundaryin Oregon. Journal of Geology 110, 687–696.

Sheldon, N.D., Tabor, N.J., 2009. Quantitative paleoenvironmental and paleoclimaticreconstruction using paleosols. Earth Science Reviews 95, 1–95.

Snigirevskaya, I.S., 1984. Kornev'ie sistemy arkheopterisov'kh v verkhnem deboneDonbassa (Root systems of archaeopterids from the Upper Devonian of Donbas).Ezegodnik Vsesusnogo Paleontologicheskogo Obsvesta Leningrad 27, 28–41.

Soil Survey Staff, 2001. Keys to Soil Taxonomy. Pocahontas Press, Blacksburg. 600 pp.Sparling, D.R., 1999. Conodonts from the Prout Dolomite of north-central Ohio and

Givetian (upper Middle Devonian) correlation problems. Journal of Paleontology73, 892–907.

Spenceley, A.P., 1983. Aspects of the development of mangals in the Townsville region,north Queensland, Australia. In: Teas, H.J. (Ed.), Biology and Ecology of Mangroves.Junk, Hague, pp. 47–56.

Stein, W.E., Hueber, 1989. The anatomy of Pseudosporochnus: P. hueberi from theDevonian of New York. Review of Palaeobotany and Palynology 60, 311–359.

Stein, W.E., Mannolini, F., Hernick, L.V., Landing, E., Berry, C.M., 2007. Giantcladoxylopsid trees resolve the enigma of the Earth's earliest forest stumps atGilboa. Nature 446, 904–907.

Stets, J., Schäfer, A., 2009. The Siegenian delta: land–sea transitions at the northernmargin of the Rhenohercynian Basin. In: Königshof, P. (Ed.), Devonian Change: CaseStudies in Palaeogeography and Palaeoecology: Geological Society of LondonSpecial Publication, 314, pp. 37–72.

Taylor, F.J., 1983. The New Zealand mangrove association. In: Teas, H.J. (Ed.), Biologyand Ecology of Mangroves. Junk, Hague, pp. 77–80.

Ufnar, D.F., Gröcke, D.R., Beddows, P.A., 2008. Assessing pedogenic calcite stable isotopevalues; can positive linear covariant trends be used to quantify palaeo-evaporationrates? Chemical Geology 256, 46–51.

Upchurch, G.R., Dilcher, D.L., 1990. Cenomanian angiosperm leaf megafossils, DakotaFormation, Rose Creek locality, Jefferson County, southeastern Nebraska. U.S.Geological Survey Bulletin 1915, 1–55.

van der Schalie, H., 1938. The naiad fauna of the Huron River in southeastern Michigan.Museum of Zoology University of Michigan Miscellaneous Publications 40, 1–83.

Ward, P.D., 2009. The Medea Hypothesis. Princeton University Press, Princeton. 180 pp.Ward, P.D., Labandeira, C., Maurin, M., Berner, R.A., 2006. Confirmation of Romer's Gap-

as a low oxygen interval constraining the timing of initial arthropod and



vertebrate terrestrialization. U.S. National Academy of Sciences Proceedings 103,16818–16822.

Westbrooke, M., Florentine, S., 2005. Rainfall-driven Episodic Flood Events: are they amajor factor in moulding New South Wales arid land vegetation patterns?Australian Geographer 36, 171–181.

White, D.A., 1907. A remarkable fossil tree trunk from the Middle Devonic of New York.New York State Museum Bulletin 107, 327–340.

Williams, C.A., Hills, L.V., Krause, F.F., 1996. Preserved organic matter andmiospores in aburied Middle Devonian (Givetian) paleosols: indicators of weathering oxidationand maturity. Catena 28, 1–19.

Work, D.M., Mason, C.E., Klapper, G., 2007. The Middle Devonian (Givetian) ammonoidPharciceras from the New Albany Shale, Kentucky. Journal of Paleontology 81,1510–1515.


author's personal copy€¦ · paleoenvironments at east windham (fig. 8). the manorkill...

Documents